Two Pleiotropic Classes of daf-2 Mutation Affect Larval Arrest, Adult Behavior, Reproduction and Longevity in Caenorhabditis elegans

Corresponding author: Donald L. Riddle, Molecular Biology Program, 311 Tucker Hall, University of Missouri, Columbia, MO 65211., riddled{at}missouri.edu (E-mail).

Communicating editor: I. GREENWALD

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The nematode Caenorhabditis elegans responds to overcrowding and scarcity of food by arresting development as a dauer larva, a nonfeeding, long-lived, stress-resistant, alternative third-larval stage. Previous work has shown that mutations in the genes daf-2 (encoding a member of the insulin receptor family) and age-1 (encoding a PI 3-kinase) result in constitutive formation of dauer larvae (Daf-c), increased adult longevity (Age), and increased intrinsic thermotolerance (Itt). Some daf-2 mutants have additional developmental, behavioral, and reproductive defects. We have characterized in detail 15 temperature-sensitive and 1 nonconditional daf-2 allele to investigate the extent of daf-2 mutant defects and to examine whether specific mutant traits correlate with each other. The greatest longevity seen in daf-2 mutant adults was approximately three times that of wild type. The temperature-sensitive daf-2 mutants fell into two overlapping classes, including eight class 1 mutants, which are Daf-c, Age, and Itt, and exhibit low levels of L1 arrest at 25.5°. Seven class 2 mutants also exhibit the class 1 defects as well as some or all of the following: reduced adult motility, abnormal adult body and gonad morphology, high levels of embryonic and L1 arrest, production of progeny late in life, and reduced brood size. The strengths of the Daf-c, Age, and Itt phenotypes largely correlated with each other but not with the strength of class 2-specific defects. This suggests that the DAF-2 receptor is bifunctional. Examination of the null phenotype revealed a maternally rescued egg, L1 lethal component, and a nonconditional Daf-c component. With respect to the Daf-c phenotype, the dauer-defective (Daf-d) mutation daf-12(m20) was epistatic to daf-2 class 1 alleles but not the severe class 2 alleles tested. All daf-2 mutant defects were suppressed by the daf-d mutation daf-16(m26). Our findings suggest a new model for daf-2, age-1, daf-12, and daf-16 interactions.

THUS far, C. elegans is the only metazoan organism in which a number of single gene mutations causing large increases in life span have been identified. These genes include daf-2 (KENYON et al. 1993 ) and age-1 (FRIEDMAN and JOHNSON 1988 ), formerly also known as daf-23 (MALONE et al. 1996 ; MORRIS et al. 1996 ; TISSENBAUM and RUVKUN 1998 ). These mutations result in mean life spans of up to 250 and 300% of wild type, respectively (LARSEN et al. 1995 ). Both genes also control dauer larva formation (RIDDLE 1988 ; GOTTLIEB and RUVKUN 1994 ).

Dauer larvae are nonfeeding, developmentally arrested, alternative third-stage larvae, which form in response to crowding and reduced food supply (CASSADA and RUSSELL 1975 ; GOLDEN and RIDDLE 1984A ). A constitutively released dauer-inducing pheromone serves as a measure of population density. A low pheromone:food ratio and low temperature promote continuous development through four larval stages (L1–L4) to the adult, but high pheromone levels and higher temperatures promote dauer formation and inhibit exit from the dauer state. Under the latter conditions, L1 larvae molt to a predauer (d) L2d stage, which lasts longer than the L2 (GOLDEN and RIDDLE 1984A ). L2ds retain the developmental potential to molt to the L3 should conditions improve, but if they do not, they molt into the dauer stage, shrink radially, and become resistant to detergent treatment and other environmental insults.

Dauer larvae are long lived relative to the adult, with maximum life spans of around 70 days (KLASS and HIRSH 1976 ) and 30 days (JOHNSON and WOOD 1982 ), respectively, in liquid culture. Dauer larvae are considered nonaging because the length of time spent in the dauer stage has no effect on postdauer life span (KLASS and HIRSH 1976 ). The basis for the enhanced longevity of dauer larvae is unknown. However, evidence suggests that a reduction in metabolic activity occurs, consistent with long-term survival in the absence of food (O'RIORDAN and BURNELL 1989 ; WADSWORTH and RIDDLE 1989 ), and although capable of rapid movement, dauer larvae are largely inactive. Dauer longevity may also be enhanced by increased resistance to stress. Dauer larvae show enhanced resistance to thermal injury (ANDERSON 1978 ) and to oxidative damage-inducing chemicals (LARSEN 1993 ). They also show increased activities of superoxide dismutase (ANDERSON 1982 ; LARSEN 1993 ) and catalase (VANFLETEREN and DE VREESE 1995 ) relative to adults.

Over 30 genes controlling dauer larva formation have been identified. Mutations in these daf genes result in either the inability to form dauer larvae in response to crowding and starvation (dauer-defective, or Daf-d), or the constitutive formation of dauer larvae in the presence of abundant food (dauer-constitutive, or Daf-c). Studies of the phenotypes resulting from combinations of daf-c and daf-d mutations have allowed the daf genes to be ordered into complex, branched pathways (reviewed by RIDDLE and ALBERT 1997 ). Mutations in one branch of the pathway (age-1, daf-2, daf-16, and daf-18) affect both dauer larva formation and adult life span (KENYON et al. 1993 ; LARSEN et al. 1995 ; DORMAN et al. 1995 ). All daf-2 and most age-1 mutants are Daf-c and may be involved in transduction of environmental information via the nervous system or by some other route. The age-1 gene encodes a putative phosphatidylinositol 3-OH kinase catalytic subunit (MORRIS et al. 1996 ). PI 3-kinases typically transmit signals from cell-surface receptor tyrosine kinases into the cell (KAPELLER and CANTLEY 1994 ). The daf-2 gene encodes a receptor tyrosine kinase similar to the vertebrate and Drosophila insulin receptors (KIMURA et al. 1997 ).

Adult expression of functions normally expressed in the dauer stage may account for the increased longevity (Age) of daf-2 (KENYON et al. 1993 ) and age-1 adults. The Age phenotype requires daf-16 activity because a daf-d mutation in this gene suppresses the enhanced longevity resulting from mutations in daf-2 (KENYON et al. 1993 ) or age-1 (LARSEN et al. 1995 ; DORMAN et al. 1995 ). The daf-16 gene encodes a Fork head-related transcription factor (OGG et al. 1997 ; LIN et al. 1997 ). The establishment of a causal link between misexpression of a particular dauer trait in the adult and extension of adult life span would be illuminating with respect to the nature of the biological determinants of life span.

Previous comparisons between the canonical allele, daf-2(e1370), and other daf-2 alleles have revealed considerable variation in the daf-2 mutant phenotype. For example, the phenotypes of daf-2(e1370) and daf-2(m41) differ with respect to temperature dependence of the Age phenotype, effects on fecundity, and interactions with mutations in the daf-d gene daf-12 (LARSEN et al. 1995 ). The daf-2(e979) mutation results in embryonic and L1 arrest (VOWELS and THOMAS 1992 ).

To better understand daf-2 function, we have conducted a detailed phenotypic study of 16 mutant alleles. We have focused, in particular, upon the range of effects on life span, how the severity of other mutant traits correlates with the degree of life span extension, and how different daf-2 alleles interact with daf-12 to affect both dauer larva formation and life span. Our results reveal the existence of two overlapping classes of daf-2 allele, differing both in their mutant phenotypes and in their interactions with daf-12. Class 1 mutants display Daf-c, Age, and increased intrinsic thermotolerance (Itt) phenotypes. Class 2 mutants are more pleiotropic, exhibiting the class 1 defects, in addition to other developmental and behavioral defects, and resembling severe age-1 mutants with respect to their larval arrest phenotype and interactions with daf-12 (GOTTLIEB and RUVKUN 1994 ; LARSEN et al. 1995 ). Defects resulting from both mutant classes are suppressed by daf-d mutations in daf-16, whereas mutations in daf-12 suppress the class 1 Daf-c and Age phenotype but not the class 2 mutants we tested. Our findings clarify the relationship between daf-2, age-1, daf-12, and daf-16 in the genetic pathway controlling dauer formation and life span.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Culture methods and strains:
Animals were maintained monoxenically in 60-mm Petri dishes containing 10 ml NG agar seeded with Escherichia coli OP50 as the food source (BRENNER 1974 ). The daf-2 mutations used in this study were e979, e1365, e1368, e1369, e1370, e1371, e1391, m41, m65, m120, m212, m577, m579, m596, m631, m632, sa193, and sa223. Other mutations used were LG I, daf-16(m26); LG III, dpy-1(e1), mec-12(e1605), unc-32(e189), unc-93(e1500); LG V, dpy-11(e224); and LG X, daf-12(m20). The daf-2(sa193) and daf-2(sa223) strains were provided by J. H. THOMAS. In VOWELS and THOMAS 1992 , daf-2(e979) was referred to as daf-2(e1286) (J. H. THOMAS, personal communication).

All alleles were backcrossed to the Caenorhabditis Genetics Center wild-type (N2) male stock at least three times to remove possible second-site mutations. Because most strains had previously been backcrossed once or twice, it was necessary to perform one or two further backcrosses. The twice backcrossed sa223 strain was backcrossed two more times only after its detailed characterization, but tests on dauer larva formation and life span indicated that the four-times backcrossed strain was indistinguishable from the strain originally received.

Construction of daf-2; daf-12 double mutants:
In the case of ts daf-2 alleles, daf-2; daf-12 double mutants were constructed as previously described (LARSEN et al. 1995 ). The daf-2(m65) III mutation results in nonconditional dauer larva formation. Consequently, the construction of the double mutant carrying daf-12(m20) X used qC1 [dpy-19(e1259ts) glp-1(q339)] III to balance m65. daf-2(m65)/qC1 males were mated with daf-12(m20) hermaphrodites at 20°, and the F₁ males were backcrossed with balanced daf-2 hermaphrodites. F₁ hermaphrodites were selfed individually at 15° to identify cross progeny of genotype m65/qC1; m20/+ based on the segregation of dauer larvae (m65), sterile adults defective in germline proliferation (Glp) (qC1), and longer animals (some m20 homozygotes are longer than wild type). Also observed among the segregants were dark-bodied animals that exhibited a novel developmental arrest phenotype (see RESULTS). A long segregant was used to establish the daf-2(m65)/qC1; daf-12(m20) strain.

Construction of daf-2(m65) unc-32(e189)/qC1 strain:
dpy-1(e1) unc-32(e189)/++ males were crossed with daf-2(m65)/qC1 hermaphrodites, and F₁ hermaphrodites were selfed. Two of 12 F₁s segregated dauer larvae and Dpy Unc (dumpy, short body; uncoordinated movement) animals. Ten Unc non-Dpy F₂s were selfed, seven of which segregated Unc dauer larvae and Dpy Uncs (i.e., were + daf-2 unc-32/dpy-1 + unc-32). Unc non-Dpy F₃ hermaphrodites were crossed with N2 males, progeny males were crossed with daf-2(m65)/qC1, hermaphrodite progeny were selfed, and a daf-2 unc-32/qC1 strain was identified as one segregating Unc dauer and Dpy Glp progeny at 20°.

Construction of heteroallelic strains:
To test the possible dominance of daf-2(e979), a daf-2(e979)/daf-2(m577) strain was constructed as follows. e979/+ males were crossed with m577 hermaphrodites at 22.5°, at which temperature m577 homozygotes do not form dauer larvae. Dauer progeny were picked and allowed to recover at 15°. Approximately half proved to be male, confirming that such dauer progeny resulted from crossing. For progeny testing at 25.5°, dauer larvae that recovered to adult hermaphrodites were allowed to lay eggs for 24 hr, then removed. Progeny were scored after 72 hr of development (measured from the middle of the 24-hr egg-laying period).

Animals heteroallelic for different combinations of ts daf-2 alleles and nonconditional daf-2 alleles were constructed to examine the 25° phenotype of nonconditional segregants in the absence (or severe reduction) of maternal rescue. To construct daf-2(m65) unc-32(e189)/daf-2(m577) +, daf-2(m65) unc-32/qC1 males were crossed with daf-2(m577) hermaphrodites at 22.5°. Since m577 does not result in dauer formation at 22.5°, dauer progeny were necessarily daf-2(m65) unc-32/daf-2(m577) +. They were induced to resume development by transfer to 15°.

The phenotypes of daf-2(m65)/daf-2(e979), mDf12/daf-2(e979), and mDf12/mDf11 at 25.5° were determined by scoring brooded cross plates 2, 3 and, if necessary, 4 days after the parental animals were transferred to fresh plates. daf-2(m65)qC1 or mDf12/qC1 males were mated to either daf-2(e979) or mDf11/qC1 hermaphrodites at 25.5°. The number of self-progeny was minimized by mating 10 adult males with two L4 hermaphrodites on plates with a 1-cm-diameter spot of bacteria.

Dauer formation, predauer arrest, and brood size assays:
The effects of daf-2 mutations on brood size, dauer formation, and early larval arrest were examined at 15°, 20°, 22.5°, and 25°. Gravid adults (10–20) grown at 15° were allowed to lay eggs for 1 hr, then removed. The resulting synchronous population was raised at 15° until the late L4 stage. Ten animals were placed singly on plates and shifted to the assay temperature. These P₀ animals were transferred to new plates every 24 hr until the end of the reproductive period. Each brood plate was examined daily to follow development to a terminal phenotype. Any adults or L4 larvae were counted and removed. The number of dauer progeny was scored 72 hr after the midpoint of egg laying at 25.5°, 80 hr at 22.5°, 96 hr at 20°, and 120 hr at 15°. Samples compromised by fungal or bacterial contaminants were excluded. In the case of daf-2(sa223), maternally rescued adults were picked from among progeny of daf-2(sa223)/qC1 hermaphrodites. Although these sa223 progeny often arrested development as L3s, L4s, or adults (exhibiting darkness of body, reduced motility, and reduced gonad development), a number of adult sa223 homozygotes developed gonads, and these were picked for brood-size analysis.

In studies of larval development of daf-2; daf-12 strains, groups of gravid hermaphrodites raised at 15° were allowed to lay eggs overnight (5–6 animals, 22.5°) or for 4 hr (10–15 animals, 25.5°), then removed. At 22.5°, the appearance of larvae was checked 60 hr after the midpoint of egg laying and scored after 80 hr. At 25.5°, the appearance of larvae was checked after 50 hr. Progeny were intermittently observed and scored between 50 and 100 hr after the midpoint of egg laying.

Life span determination:
L4 larvae grown at 15° were placed at 15° and 22.5°, typically at a density of 15–30 animals per plate. These were transferred daily to fresh plates during the egg-laying period and subsequently at approximately weekly intervals. Death was scored as the absence of any movement and failure to move at all after several light pokes with a platinum wire. The zero time point was the time of L4 transfer. Samples compromised by bacterial contaminants were excluded. Life span was assayed at 22.5° rather than the usual nonpermissive temperature of 25.5°, at which a high level of mortality is seen throughout adult life in some daf-2 mutants due to internal hatching of eggs and other unknown causes that may not be related to senescence. It was expected that at 22.5°, population survival curves would be more rectangular and show less variation between trials, facilitating comparisons between strains.

Intrinsic thermotolerance assays:
Young adult hermaphrodites grown at 15° were transferred to prewarmed (35°) 60-mm NG agar plates (not spread with bacteria) and maintained at that temperature. The number of worms dead and alive was recorded at 2-hr intervals until all were dead. Any worms that died as the result of crawling up the wall of the plate were excluded from the analysis.

Intrinsic thermotolerance of selected daf-2(class 2); daf-12 double mutants was determined in the same manner. However, to test the effect of daf-12 on certain class 1 alleles, the procedure was modified to increase the thermotolerance of the daf-2 adults. Synchronous populations were raised to adult at 20°, then transferred to fresh plates at 25° for 2 days before testing, as described above. To compensate for the slower development of the class 1 m41 mutant and the class 2 e1391 mutant, populations of strains containing these alleles were started 1 day earlier.

Adult behavior and morphology:
During the course of life span assays, behavior and appearance of adult animals were examined at 1- to 3-day intervals through a dissecting microscope. At higher temperature several alleles resulted in some shrinkage of the adult body, clearly discernible at x25 magnification, and gonadal abnormalities, which were easily visible at x50 magnification as clear regions against the otherwise dark body characteristic of daf-2 adults. The onset of an obvious reduction in motility and the appearance of coiling behavior was generally rapid (occurring over a 1- to 2-day period), such that motility was readily classified as normal or reduced. Reproducibility of scoring was confirmed by consistency in classification in blind trails on successive days and by independent classification by two observers.

Pharyngeal pumping rate:
Worms were raised at 15°, then transferred singly at the L4 stage to fresh plates at 22.5°. The mean pumping rate did not include nonpumping animals. Pumping was scored over a 15-sec interval or at 30-sec or 1-min intervals where pumping rate was reduced. In graphs of pumping rate as a function of percentage of maximum life span, the latter was calculated by dividing the ages at which pumping rate was measured by maximum life spans (shown in Table 3).

Male mating efficiency:
Six daf-2 alleles were tested for their effects on male fertility at 20° and 25.5°. These alleles (e1370, e1371, e1391, m41, m120, and m577) were selected as a representative sample of the range of severity and variation in mutant phenotypes. Male stocks were established with males obtained by heat shock (SULSTON and HODGKIN 1988 ) or from males occurring spontaneously in hermaphrodite populations maintained at 15°. A standard quantitative mating test was employed (HODGKIN 1983 ), in which six late L4 daf-2 males (raised at 15°) and six late L4 dpy-11(e224) hermaphrodites were placed together on a 60-mm plate spread with bacteria and incubated for 24 hr at 20°, after which the males were removed. Total cross-progeny (non-Dpy F₁) were counted. Tests of mating efficiency at 25.5° were performed in a manner similar to those at 20° except that early (rather than late) L4 males were used, such that the entire period of spermatogenesis occurred at the higher temperature.

Isolation and characterization of daf-2 deficiencies:
Two -radiation-induced deficiencies were isolated in a noncomplementation screen. Mixed-stage N2 populations containing many males were exposed to 1500 R. Irradiated young adult males were immediately crossed to dpy-1(e1) daf-2(e1370) unc-32(e189) hermaphrodites at 20° then shifted to 25° after 24 hr. Three days later the F₁ progeny were screened by visual inspection and SDS selection for the presence of wild-type or Dpy dauer larvae. Four of seven wild-type larvae that failed to recover spontaneously at 25° recovered to the L4 stage after 2 days at 15°, and these four were crossed individually with daf-2(m65) unc-32(e189)/qC1[dpy-19(e1259) glp-1(q339)] males at 20°. All four crosses were shifted to 25° after egg laying began (approximately 24 hr). Two of the crosses gave progeny. From each, wild-type nondauer L4 hermaphrodites of putative genotype mDf[daf-2]/qC1 or dpy-1 daf-2 unc-32/qC1 were placed individually on plates. Those issuing Dpy Unc progeny were discarded. The remainder for both isolates gave only wild-type and qC1 progeny, indicating that the new mutations were lethal when homozygous. The two putative deficiencies were named mDf11 and mDf12. Complementation testing showed that neither deficiency deleted unc-93; tests against mec-12 were inconclusive. Thus, the putative deficiency endpoints lie between dpy-1 and unc-93. Progeny counts confirmed that both strains gave approximately 25% embryonic lethal progeny. The mDf12/qC1 animals, but not mDf11/qC1, grow slowly relative to qC1 homozygotes.

Both deficiencies were shown by PCR analysis to lack daf-2 sequences encoding portions of the extracellular domain and the tyrosine kinase domain. The sequence of the daf-2 cDNA was obtained from GenBank (accession no. AF012437) and compared to the C. elegans sequence database to identify genomic YAC clones corresponding to daf-2 (KIMURA et al. 1997 ). Two pairs of oligonucleotide primers were designed for PCR, one pair each from genomic sequence encoding the extracellular and intracellular domains of the protein. Primers CTCTCGAACAAAACAGTGCCTATC and AATGAGGGCCAACTAAAGAAGACC amplified a 659-bp wild-type fragment encoding a portion of the extracellular domain, whereas primers TTCGGACCGTGTGCTATTAAGATT and CTCGGACCTCCACTATGATTCATC amplified a 1082-bp fragment encoding a portion of the kinase domain. Another primer pair (AGCAGCACCAGCAACAGGAGTAAC and TTTCAAACCCCCAACTCATACCTC) from the lin-31 region of chromosome II (cosmid K10G6) was used as an internal positive control to confirm that amplifiable DNA from the deficiency homozygotes was present in the reaction. This primer pair amplified a 523-bp product.

To identify and isolate deficiency homozygotes (dead eggs), newly starved plates of mDf/qC1 bearing large numbers of unhatched eggs were washed free of gravid adults and most larvae by two gentle rinses with sterile M9 buffer, leaving most eggs still adhering to the agar surface. Washed plates were incubated at 20° for 24 hr and washed again to remove larvae that hatched after the initial rinse. Deficiency homozygotes were identified as eggs that appeared abnormal in shape and remained unhatched after a further 24–48 hr of incubation at 20°. These were picked individually or in groups of up to 15 using a pulled-out 20-µl pipette filled with chitinase solution (WILLIAMS et al. 1992 ).

PCR reactions (25 µl final volume) were performed according to WILLIAMS et al. 1992 , except that Taq polymerase (Fisher Scientific) was used at 0.5 unit per reaction, and "master mix" was added to each reaction in 18.5-µl volumes to allow for separate addition of daf-2 and lin-31 control primers. Test reactions on N2 DNA using a mix of control and daf-2 primers resulted in production of both predicted products, whereas parallel reactions with putative Df DNA gave only the control product. For each strain and primer pair, assays were performed in duplicate on purified N2 DNA, deionized water blanks, and worm extracts (Df/qC1 heterozygotes) and in quadruplicate on eggs (Df homozygotes). All reactions were brought to 95° rapidly and held for 3 min, cycled 30 or 50 times (95°/30 sec, 58°/30 sec, 72°/60 sec), then held for 7 min at 72°. Amplification products were resolved on 1–2% agarose minigels.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Selection of alleles for study:
Most previous studies of daf-2 used the canonical allele, e1370. We selected 15 additional alleles to cover a wide range of severity of the Daf-c phenotype. Fourteen alleles, including e1370, were selected from the 40 currently in the Riddle lab collection (of which 28 are conditional and 12 nonconditional). Two other alleles, sa193 and sa223, were provided by J. H. THOMAS. All are recessive. Some phenotypic characterization had previously been reported for 8 of the 16 alleles (including 2, m65 and sa223, that are recessive lethal and are maintained in heterozygous stocks) as listed in Table 1. The remaining 8 are previously undescribed, temperature-sensitive (ts) Daf-c mutants and were isolated from mutant screens over the last 25 years in the MRC Laboratory of Molecular Biology, Cambridge, United Kingdom (1973–1975) or subsequently in the Riddle laboratory. All alleles were backcrossed to N2 at least three times to avoid possible phenotypic variation due to differences in genetic background.

Characterization of alleles:
We focused on the phenotypic effects of hypomorphic daf-2 alleles, although analysis of daf-2(e979), nonconditional daf-2 alleles, and daf-2 deficiencies suggested that the null [daf-2(0)] phenotype has an embryonic and L1 arrest component as well as a nonconditional Daf-c component (see below). By focusing on weaker alleles we have been able to examine the role of daf-2 in later larval development and in the biology of the adult.

Examination of the phenotypes of 16 daf-2 mutants revealed two types of daf-2 allele, class 1 and class 2. The following sections on the phenotypic analysis begin with descriptions of the major traits common to the class 1 and class 2 alleles: Daf-c, Age, and Itt. Next, L1 arrest is described, which results from all daf-2 alleles but at a greatly elevated frequency in two class 2 alleles. Then follow descriptions of the diverse class 2-specific traits, with an overview and more refined classification of the 16 alleles.

Constitutive dauer larva formation:
The percentage of the mutant population that constitutively entered the dauer stage in abundant food was measured at 15°, 20°, 22.5°, and 25.5° for all ts alleles (Table 2). These alleles largely formed a continuous series in severity, from the weakest, e1365 and m577, to the strongest, e1369. Although the five weakest alleles with respect to Daf-c are class 1, some class 1 alleles exhibited a stronger Daf-c phenotype than most class 2 alleles. Differences in the severity of most alleles were revealed at 22.5°; the majority formed no dauer larvae at 15° but formed all dauer larvae at 25.5°.

At 22.5° most alleles resulted in the rapid development of larvae either into L3 or dauer larvae (either transient or persistent). The sa193 and e1371 alleles were ranked in severity on the basis of the percentage of dauer larvae that recovered spontaneously within 80 hr after eggs were laid (Table 2). However, in two instances (e1370 and m579) all larvae arrested initially at the predauer L2d stage, followed by slow development either into dauer larvae or into stages apparently intermediate between the L2d and L3 or L4 stages. These dauer-like animals resembled L2ds in their darkness of body yet resembled L3s or L4s in size and gonadal development (Figure 1B).

View larger version (39K): In this window In a new window Download PPT slide   Figure 1. Morphology of daf-2 mutant larvae and adults. (A) daf-2(e1370) L3, raised in abundant food at 15°; (B) e1370 dauer-like L3, raised at 22.5°; (C) 5-day-old e1370 hermaphrodite maintained at 15°; (D) daf-2(sa193) hermaphrodite transferred to 25.5° at the L4 stage and incubated for 3 days; (E) e1370 hermaphrodite transferred to 25.5° at the L4 stage and incubated for 3 days; (F) e1370 dauer larva raised at 25.5°. (A–F) Scale bar, 0.2 mm.

Two exceptional alleles were e979, which at 25.5° arrested as embryos, L1s, or rarely at a stage resembling L2d larvae (as previously reported by VOWELS and THOMAS 1992 ), and sa223, which formed dauer larvae at 25.5° but at lower temperatures largely arrested development as L2ds (MALONE and THOMAS 1994 ). At 25.5°, three of the daf-2 mutants (e1368, e1371, and sa193) formed dauer larvae that displayed sporadic pharyngeal movement resembling the pumping of feeding animals. Normal dauer larvae do not feed, the buccal cavity is closed, and pharyngeal pumping is not observed.

To examine maintenance of the dauer stage, dauer larvae from all ts Daf-c strains were maintained at 25.5° for 8 days after hatching and examined daily for the presence of adults. A high proportion of e1371 and sa193 dauer larvae resumed development. Occasional adults were seen among e1368 and m212 dauer larva populations (all population sizes >200). Thus, in the cases of e1368, e1371, and sa193, pharyngeal movement in dauer larvae corresponded with some degree of recovery from developmental arrest.

To determine whether nondauer development was affected by daf-2 mutations, we tested the effect of 14 daf-2 alleles on the subsequent development of animals raised at 15° and transferred to 25.5° as early L3s. Only in the case of e1371 did all the animals grow to gravid adults within 24 hr. In most other cases, the majority of animals developed into gravid adult hermaphrodites, with a minority (5–20%) of pregravid adults. In three cases, e979, e1370, and e1391, abnormal development occurred; 24 hr after upshift, most animals were abnormally dark and thin, with underdeveloped gonads, and in many cases a protruding vulva. After 48 hr, some gonadal development was observed in most of the e979 and e1391 adults and in all of the e1370 adults.

Hermaphrodite life span:
The effect of 15 daf-2 alleles on adult hermaphrodite life span was measured at 15° and 22.5° (Figure 2, Table 3). All mutations increase adult life span. The alleles exhibited a gradient of increasing longevity at 15°, but the relative extensions of median and maximum life span were similar in most cases (Figure 2A, Table 3). The greatest increases in maximum life span were approximately 300% of the N2 control. At 15°, increases in life span were only marginal for e1365, m41, and m577.

View larger version (98K): In this window In a new window Download PPT slide   Figure 2. Effect of daf-2 mutations on median and maximum hermaphrodite life span. Median and maximum life span are expressed as a percentage of that of the wild-type control. Bars represent standard error. (A) Life span at 15°; (B) life span at 22.5°. Alleles are in order of increasing severity of the L2d and dauer arrest phenotype within each class.

At 22.5° all 15 mutants had clearly extended life spans (Figure 2B, Table 3). Most weaker alleles showed greater extensions in life span relative to N2 at 22.5° than at 15°. However, the five alleles with the largest increases in median and maximum life span at 15°, e979, e1369, e1391, m212, and sa223, at 22.5° generally showed similar or reduced extension of median life span and similar extensions in maximum life span (Figure 2, Table 3). The greatest percentage increases in life span at 22.5° did not exceed those seen at 15°, suggesting that the limit of adult life span extension that may result from loss of daf-2(+) function has been reached in these cases. At either temperature, this longevity ceiling represents a 2.5-fold increase in median life span and a 3-fold increase in maximum life span relative to wild type.

In some class 2 alleles, the extension of maximum life span at 22.5° greatly exceeded that of median life span (Figure 2B). This suggests either that these mutations have a deleterious effect on adults, resulting in some premature deaths, or conceivably that they cause individuals in the population to age at different rates.

Intrinsic thermotolerance:
Mutations in daf-2 and age-1 also result in increased tolerance to thermal stress, and it has been suggested that increased Itt, as measured by time of survival at 35°, is a necessary condition for the Age phenotype (LITHGOW et al. 1995 ). Because mutant daf-2 alleles result in a range of Age phenotypes at 15°, the Itt phenotype of populations of young adults raised to the age of onset of egg laying at 15° was tested (Figure 3, Table 4). Increased Itt was observed in all alleles tested except m41 and m577. Seven of the most severely Age alleles at 15° (e979, e1369, e1370, e1391, m212, m579, and m596) showed similarly high values for Itt, suggesting that, as in the case of life span, the limit of adult thermotolerance that can result from loss of daf-2(+) function has been reached in these cases. The severely Age allele sa223 was also strongly Itt (data not shown). This suggests that loss of daf-2 function in mutants that can grow to the adult at 15° may result in, at most, approximately a 75% increase in median survival at 35° (Table 4).

View larger version (18K): In this window In a new window Download PPT slide   Figure 3. Intrinsic thermotolerance assays. A representative sample of 35° survival curves for daf-2 adult hermaphrodite populations, raised at 15°, is shown.

Embryonic and L1 arrest:
daf-2(e979) exhibits almost 100% embryonic or L1 arrest at 25°. This phenotype was examined in all 15 ts alleles at 15°, 20°, and 25.5° (Table 5). Some embryonic and L1 arrest (mostly the latter), higher than that observed in N2 controls, was observed in all daf-2 alleles tested at 25.5° (except sa193) but in most cases not at 15° or 20° (Table 5). The frequency of such arrest was generally low or undetectable at the temperatures tested (mean, <6%). However, at 25.5° the e1391 and e979 mutations resulted in 10.5 ± 2.8 and 100% embryonic and L1 arrest, respectively. A high level of sa223 embryonic arrest (14/29 total progeny) was seen at 15°.

Brood size:
Fecundity of daf-2 mutant hermaphrodites was assayed at 15°, 20°, 22.5°, and 25.5° (Table 6). In self-fertilizing Caenorhabditis elegans hermaphrodites, maximum brood size normally reflects the fixed number of sperm produced before the switch from spermatogenesis to oogenesis. When averaged over all temperatures tested, class 1 mutants have broods ranging from 85 to 100% that of N2, and class 2 mutants (except sa223) have broods ranging from 60 to 93% of N2. Two class 1 alleles (e1368 and e1369) were cold sensitive, showing 35–40% reductions in brood size only at 15°. These were the only class 1 alleles where significant brood size reductions (P < 0.01) were seen, whereas most class 2 alleles resulted in significant reductions in brood size, which were generally more severe at higher temperatures (Table 6).

The largest reductions in brood size resulted from sa223. In this case, brood sizes of maternally rescued animals were measured because progeny of sa223 homozygotes do not develop to adulthood. Mean sa223 brood size did not exceed 4% of wild type at any of the four temperatures tested (Table 6). At all temperatures, some sa223 adults were sterile, even though adults in which some gonad development was evident were selected (see MATERIALS AND METHODS).

Late progeny:
Production of progeny by N2 hermaphrodites at 25.5° ceases after 3–5 days, but e1370 hermaphrodites continue to produce occasional progeny for at least 50 days (LARSEN et al. 1995 ). All 15 ts daf-2 alleles were examined for late progeny production at 15°, 22.5°, and 25.5° (Table 7). Of these, only the 7 class 2 alleles resulted in this trait.

Because late progeny generally appeared after the transfer of hermaphrodites to fresh plates, and often produced in bursts by a single individual, it is possible that a cue present in fresh bacterial lawns might stimulate late reproduction in these mutants. At least some late progeny were released by their mothers as L1 larvae rather than as eggs. Late eggs were never observed.

Late progeny could result from a delay in production of oocytes or from renewal of spermatogenesis later in life, after depletion of the sperm produced during the L4 stage. To test this latter possibility, the laying of unfertilized oocytes at the end of the egg-laying period, an indicator of sperm depletion (WARD and CARREL 1979 ), was monitored at 22.5° (Table 8). All daf-2 mutant alleles resulted in a reduction in the number of oocytes laid relative to the N2 control, although this reduction was not significant in many class 1 alleles. The seven class 2 mutants, which exhibited late progeny production, also laid the fewest oocytes. Hence, stored sperm may be present in hermaphrodites of the late progeny-producing strains at the end of the normal egg-laying period, and late progeny may result from fertilization of oocytes that are not produced until late in life. That daf-2(e1370) hermaphrodites are defective in oocyte production is also suggested by the abnormal appearance of the gonad. Microscopic examination with Nomarski optics of adult daf-2(e1370) hermaphrodites 19 hr after shifting from 15° to 25.5° revealed that, instead of the usual 8–10 oocytes stacked in the proximal arm and loop of each gonad arm, only 2–4 oocytes were present (J. MCCARTER, personal communication).

Maternal rescue of larval arrest:
To test for maternal rescue of the Daf-c trait, hermaphrodites from the 15 ts daf-2 strains were crossed with N2 males at 25.5°, and F₁ heterozygotes were selfed at the same temperature. Their progeny were scored after 48 hr. Because at 25.5° all, or almost all, homozygous daf-2 mutants form dauer larvae, 25% of F₂ animals were expected to form dauer larvae in the absence of maternal rescue. The ratio of nondauer-to-dauer larva progeny was approximately 3:1, except for m41 heterozygotes, which segregated only 12 ± 8% dauer larvae, indicating that the Daf-c phenotype was maternally rescued in approximately half of the m41 homozygous progeny (Table 9; data not shown for ts alleles where no maternal rescue was seen). Repeating the m41 test at 25.5°, + m41 +/ dpy-1 + unc-32 parents issued 8% dauer larvae, 66% wild-type L4-adult and 25% DpyUnc progeny (N = 185). The e979 embryonic and L1 arrest phenotype was found to be fully maternally rescued such that e979 homozygotes arrested development at the dauer stage or, occasionally, the L2d stage (Table 9).

The m65 mutation results in nonconditional dauer larva formation, and sa223 results in formation of L2ds, dauer larvae, and sterile or near-sterile adults. Both these alleles are maintained in trans to the balancer chromosome qC1[dpy-19(e1259) glp-1(q339)]III. No maternal rescue of m65 was detected at 25.5° when assayed by scoring the progeny of m65/qC1 hermaphrodites (Table 9). Homozygous sa223 progeny of sa223/qC1 either arrested permanently at a stage resembling late L2d or developed to adulthood via dauer-like third and fourth stages. These adults were sometimes thin and dark bodied, and when gonads were present, they often appeared abnormal.

Comparison of progeny of homozygous sa223 and heterozygous sa223/qC1 hermaphrodites (Table 2 and Table 9, respectively) revealed maternal rescue of the L2d-like arrest but no rescue of the Daf-c phenotype. Animals homozygous for the sa223 allele formed dauer larvae at 25.5° but very few at 22.5°, whether they were progeny of sa223 homozygotes (Table 2) or of sa223/qC1 (Table 9). These results suggest that, whereas the Daf-c phenotype is not normally maternally rescued (with the exception of m41), the L2d arrest phenotype resulting from certain alleles may be.

Death from internal hatching (matricide):
Animals that died as the consequence of internal hatching of eggs were excluded from measurements of life span. In most daf-2 mutants, matricide, resulting in a "bag of worms" phenotype (TRENT et al. 1983 ), was observed at a low frequency (<20%), occurring more readily at 22.5° than 15° (data not shown). At 22.5°, a higher frequency of matricide (over 30%) resulted from four class 2 alleles: e979, e1370, e1391, and sa223.

Adult morphology and behavior:
Certain daf-2 alleles (e.g., e1370) resulted in progressive changes in adult hermaphrodite appearance and behavior during the 3 days after shifting L4s from 15° to 22.5° or 25.5°. Such changes were generally not seen at 15° (Figure 1C), and some alleles (e.g., sa193) did not produce such changes at any temperature (Figure 1D). One day after transfer, e1370 adults were egg-laying defective (Egl), their bodies swollen with retained eggs. Over subsequent days, overall body darkening and shrinkage in diameter were seen. The gonad and intestine, viewed in the stereomicroscope at x50 magnification, appeared to become squeezed by the contracting body, the intestine becoming reduced to a narrow, longitudinal strip. In such animals, the gonad had an abnormal, clear appearance (Figure 1E). These animals also displayed an uncoordinated phenotype, comprising reduced movement, coiling behavior, and frequent adoption of a kinked posture similar to that seen in dauer larvae (Figure 1E and Figure F).

Appearance and behavior of the 15 ts daf-2 mutants was monitored during the first 2 weeks of adulthood at 15°, 22.5°, and 25.5°. Morphological and behavioral abnormalities similar to those seen in e1370 mutants resulted from 11 of the 15 alleles in a temperature-sensitive manner (Table 10), whereas sa223 exhibited these defects at all temperatures. At 22.5°, only 8 of the 15 alleles resulted in reduced motility. Timing and rate of onset of this Unc trait showed considerable variation among alleles (Table 10). Interestingly, in several alleles adult motility appeared normal for an extended period before becoming reduced.

The e1370 allele was tested for reversibility of the morphological and behavioral abnormalities. Twenty-six late L4 larvae that were shifted to 25.5° for 3 days became fully Unc. They were then shifted back to 15°. Two days later, wild-type motility and morphology returned, and >200 progeny were produced. Twelve control hermaphrodites of the same age left at 25.5° produced only two progeny and no unfertilized oocytes. The 24 survivors of the shift were returned to 25.5°, and by day 8 all animals were once again Unc and morphologically abnormal. Twenty survivors were again downshifted to 15°. These animals again recovered wild-type motility and appearance by day 10, at which time they were shifted up to 25.5° for the third time. By day 12, all 15 survivors were Unc and morphologically abnormal. These were shifted to 15°, and by day 13, most animals had again recovered wild-type appearance and behavior. Thus, the behavioral and morphological abnormalities displayed by e1370 hermaphrodites at higher temperatures are reversible at least three successive times in the same animals. To our knowledge, this is the only case of a reversible ts adult Unc phenotype in C. elegans.

Male fertility:
The effect of six mutant daf-2 alleles on male mating efficiency was measured using a standard quantitative mating test (HODGKIN 1983 ). Males were raised at 15° and tested at 20° and 25.5°. At 20° mating efficiency and fertility were reduced to 31–45% of wild-type controls, except for e1391, where it was reduced to 10% of wild-type controls (Table 11). At 25.5°, the class 1 (e1371, m41, and m577) males were as fertile as N2 males or even more so, whereas the class 2 males (e1370, e1391, and m120) males produced no progeny at all. The latter three alleles also resulted in greatly reduced male motility, suggesting that, as in the case of many uncoordinated mutant males (HODGKIN 1983 ), their failure to sire offspring is the result of their inability to copulate.

Pharyngeal pumping rate:
KENYON et al. 1993 compared the rate of pharyngeal pumping, an indicator of the rate of ingestion of food, in wild-type and daf-2(e1370) adult hermaphrodites at 20°. They found that although the rate of pumping declines with time at a similar rate in both strains, it is greatly reduced throughout the extended phase of the e1370 life span. Our examination of pharyngeal pumping in three class 1 and three class 2 mutants indicated that class 2 but not class 1 mutations depress the rate of pharyngeal pumping.

We examined the effect of daf-2 alleles e1371 (class 1) and e1391 (class 2) on the rate of pharyngeal pumping at 22.5° from early adulthood onward. e1391 mutants exhibited a rapid decline in the rate of pumping with advancing age (Figure 4A), whereas the rate of pumping slowed down more gradually with advancing age in e1371 animals.

View larger version (35K): In this window In a new window Download PPT slide   Figure 4. Variation of daf-2 mutant pharyngeal pumping with age at 22.5°. (A) Pharyngeal pumping rate plotted against age in days; (B) pumping rate plotted against age expressed as percentage of maximum life span; animals in which pumping was not observed in a 1-min interval were excluded from pumping rate calculations.

From day 6 onward, the rate of pumping was greater in e1371 animals than in wild type. Does this signify that e1371 results in higher rates of pumping in older animals or that e1371 animals are biologically younger at later chronological ages? To clarify this issue, pumping rate was plotted against age expressed as percentage of maximum life span (Figure 4B). The decline in pumping rate with percentage maximum life span was similar in N2 and e1371 animals; that is, the decline in pumping rate scaled with life span. This indicates that e1371 has little effect on the rate of pharyngeal pumping relative to biological or developmental (as opposed to chronological) age and also that, in e1371 mutants, pharyngeal pumping rate provides a biomarker for developmental age. Conversely, in the case of e1391, pumping is a poor indicator of developmental age. This is also true of e1370 (data not shown).

Animals in which no pumping was observed in a 1-min interval were excluded from pumping rate measurements. In the case of class 2 alleles (e1370, e1391, and m596), the majority of animals exhibited greatly reduced pumping during the first 40% of the life span, but in class 1 (e1371, m41, and sa193) mutants, as in wild-type populations, the majority of animals exhibited greatly reduced pumping only in the second half of the life span (data not shown).

Classification of daf-2 alleles:
On the basis of the above phenotypic descriptions, daf-2 alleles fell broadly into two types (Table 12). Class 1 alleles resulted in constitutive dauer larva formation (Table 2), increased longevity (Table 3), increased intrinsic thermotolerance (Table 4), a low level of embryonic and L1 arrest (mean, less than 6%) at 25.5° (Table 5), and a reduction in the number of unfertilized oocytes laid (Table 8). Class 2 mutants exhibited all of the above traits plus some or all of the following: a higher level of embryonic or L1 arrest (mean, greater than 6%) at 25.5° (Table 5), formation of dauer-like L3 and dauer-like L4 larvae (Figure 1B), reduced adult motility (Table 10), shrinkage of the adult body and gonad abnormalities at 22.5° and 25.5° (Table 10), reduced brood size (Table 6), increased frequency of internal hatching at 25.5° and production of late progeny (Table 7), greatly reduced (less than 10% that of wild type) number of oocytes laid (Table 8), and a reduction in the extension of median life span as compared to that of maximum life span (Table 3).

The class 1 alleles (except m41) can be ranked into an allelic series defined by the severity of Daf-c and Age, where the latter is expressed as maximum life span at 15° (Table 2 and Table 3). This allelic series is as follows, in order of increasing severity:

This allelic series also approximately describes the ranking in severity of Itt. The class 1 allele m41 is among the weakest alleles with respect to Age and Itt when grown at 15° (Table 3 and Table 4), whereas with respect to its Daf-c phenotype at higher temperatures it is more severe than e1368 (Table 2).

The class 1 alleles can be subdivided into four subclasses, 1A–1D (Table 12). Class 1A mutants (e1368, e1371, and sa193) do not exhibit any reduced motility at any of the temperatures tested (Table 10) and differ from other class 1 mutants in that dauer larvae display slight pharyngeal movement at 25.5° 72 hr after the egg stage and may eventually recover into adults at 25.5°. The single class 1B allele, m41, resembles a class 1A allele, except that no pharyngeal movement or dauer larva recovery occurs at 25.5°. m41 is also unique in that the Daf-c phenotype is maternally rescued, and although moderately severe with respect to Daf-c at 22.5° (Table 2), it is almost entirely ts for Age and Itt (Table 3 and Table 4). The class 1C mutants e1365, e1369, and m577 exhibit reduced motility after 7 days at 25.5° (Table 10). The single class 1D mutant, m212, exhibits reduced motility at both 22.5° and 25.5°. This allele may be considered a borderline class 2 allele.

If both class 1 and class 2 traits are considered, no consistent ranking of allele severity may be achieved. However, on the basis of the severity of the class 2-specific mutant traits alone, the class 2 alleles may be ranked in severity as

The seven class 2 alleles constitute five subclasses, 2A–2E, based on severity and extent of pleiotropy (Table 12). The class 2A allele m596 results in late progeny production only at 25.5° (Table 7), whereas the class 2B alleles m120 and m579 do so at 22.5° as well. m120 and m579 also result in slight gonad defects. In addition to these defects, the class 2C allele e1370 and all the more severe alleles in this series also result in severe gonad defects at 22.5° and 25.5°, a mean brood size of less than 60% that of wild type at 25.5° (Table 6), over 30% mean death due to internal hatching of larvae at 22.5°, and a reduction in unfertilized oocytes laid to less than 5% of wild type at 22.5° (Table 8). Furthermore, e1370 and some of the more severe alleles in the class 2 series also cause a significant reduction in median life span relative to maximum life span at 22.5° (Figure 2B). The class 2D alleles e1391 and e979 resemble e1370 but additionally result in 11 and 100% mean predauer arrest at 25.5°, respectively (Table 5). e1391 animals also do not lay unfertilized eggs at 22.5° (Table 8). Almost all of the class 2 defects are most severe in the case of sa223, which exhibits gonad defects and reduced motility even at 15° (Table 10), and a brood size reduced to less than 5% of wild type (Table 6). However, with respect to embryonic and L1 arrest, sa223 is not more severe than e979. On the whole, with respect to class 2-specific defects, alleles may be ranked such that all the defects of weaker alleles are present or more severe in stronger alleles. The overall severity ranking of class 2 alleles given above is the most parsimonious, taking all the data into consideration.

The class 2 alleles e1370, e1391, and m596 caused a reduction in the proportion of older animals pumping at 22.5°, and in the case of e1370 and e1391, a severe reduction occurred in the rate of pharyngeal pumping (Figure 4A, and Figure 4B). Conversely, the class 1 alleles e1371, m41, and sa193 had only slight effects on pharyngeal pumping. This suggests that suppression of pharyngeal pumping may be a general feature of class 2 but not class 1 alleles. Similarly, the class 2 but not the class 1 alleles tested resulted in males that could not sire progeny at 25.5° (Table 11).

Correlation between mutant traits:
Examining correlations between severities of different mutant traits may reveal which daf-2 pleiotropic effects may be manifestations of the same underlying physiological defect. With the exception of certain alleles, daf-2 mutant traits fall broadly into two clusters of apparently linked phenomena: first, Age, Daf-c, Itt, and minor traits common to all alleles; second, the class 2-specific defects. The severity of the Age/Daf-c/Itt cluster and the class 2 defects appears largely unconnected. Thus, for example, e1369 ranks among the four most severe alleles with respect to Age and Daf-c yet displays almost no class 2 defects. Of the six most severe Daf-c alleles, three are in class 1.

We examined further the correlations between Age, Daf-c, and Itt traits. Plotting maximum life span at 15° against dauer larva formation at 22.5° shows a positive correlation in severity among most alleles (Figure 5A). One exception is m41, which is short lived (at 15°) relative to the severity of its 22.5° Daf-c phenotype. The sa223 mutant is long lived relative to dauer formation at 22.5°, but it exhibits an extreme predauer arrest phenotype. Also, e1391 may be somewhat long lived relative to its Daf-c phenotype (Table 2), suggesting that, in this case, class 2 defects are associated with either an enhancement of Age or a depression of Daf-c.

View larger version (62K): In this window In a new window Download PPT slide   Figure 5. Plots of Daf-c, Age, and Itt phenotype severities, derived from data displayed in Table 2, Table 3, and Table 4, respectively. Bars represent standard errors. (A) Age versus Daf-c. Note that although sa223 results in only 3.3% dauer formation at 22.5°, the remaining 96.7% arrest development as L2ds; (B) Age versus Itt; (C) Itt versus Daf-c. Note that Itt was measured for animals grown at 15°, a temperature at which m41 is wild type.

If Age is plotted against Itt (Figure 5B), a positive correlation is also seen. When the Daf-c (22.5°) and the Itt phenotypes of animals raised at 15° were plotted against one another (Figure 5C), the nine most severe alleles with respect to Daf-c showed a similar level of Itt, with the exception of m41. This suggests that in mutants exhibiting greater than 10% dauer formation at 22.5°, Itt in animals raised at 15° is maximally penetrant.

Interactions between daf-2 and daf-12—Dauer formation:
Previous attempts to establish the epistasis relationship between daf-2 and the daf-d gene daf-12 have been complicated by differences between daf-2 alleles. Whereas daf-2(m41); daf-12(m20) animals develop into adults at 25.5°, daf-2(e1370); daf-12(m20) animals arrest development either as embryos or L1s, or near the L2 molt (YEH 1991 ; VOWELS and THOMAS 1992 ; LARSEN et al. 1995 ). To test the hypothesis that phenotypes in combination with daf-12 will be class specific, the phenotypes of six class 1 and four class 2 alleles were examined in combination with daf-12(m20). Comparing developmental phenotypes of the 10 daf-2; daf-12 strains, two components of the Daf-c phenotype, developmental arrest and dauer larva morphogenesis, may be distinguished. daf-12(m20) prevents dauer larva morphogenesis in both class 1 and 2 alleles, and it suppresses developmental arrest in class 1 but not severe class 2 alleles (Table 13). Developmental arrest was suppressed in the case of the weakest class 2 allele, m596.

The effect of each daf-2; daf-12(m20) combination on dauer formation was assayed (Table 13). In combination with daf-12(m20), all six class 1 alleles behaved similarly. At 22.5°, all animals developed into L4s or adults by 60 hr. At 25.5°, development was slightly retarded such that L3s and L4s were seen instead of adults at 50 hr, but all animals subsequently developed into adults. At 22.5°, the three daf-2; daf-12 strains containing class 2 alleles behaved similarly to one another. All permanently arrested development at a stage resembling L3 or L4 in size but with an underdeveloped gonad. At 25.5° daf-2(e1370); daf-12(m20) resulted in either embryonic or L1 arrest, or arrest at a stage apparently intermediate between an L2 or L3 and an L2d. Almost all daf-2(e1391); daf-12(m20) animals arrested as the L2- and L3-like forms seen among daf-2(e1370); daf-12(m20) populations. At 25.5°, the daf-2(e979); daf-12(m20) strain developed as in the absence of the daf-12(m20) mutation (i.e., embryonic lethality and L1 arrest).

The daf-2(m65); daf-12(m20) progeny of daf-2(m65)/qC1[dpy-19 glp-1]; daf-12(m20) animals were also examined. The m65 mutation alone results in nonconditional dauer formation. The m65; daf-12 homozygotes were developmentally arrested and resembled daf-2(e1370); daf-12 and daf-2(e1391); daf-12 arrested larvae. By 3 days at 25.5°, m65; daf-12 segregants arrested development as dark-bodied, somewhat thin, L3-sized individuals, which comprised 27 ± 3% of the total population (N = 866). At 15°, they became larger, arresting development at approximately the size of L4s. No dauer larvae were seen. These observations suggest that m65 is a class 2 allele.

Interactions between daf-2 and daf-12—Life span:
The effect of eight daf-2 alleles in combination with daf-12(m20) on adult life span was measured (Figure 6, Table 3). As in dauer formation, class 1 alleles behaved similarly to one another with respect to the effects of daf-12(m20) on the Age phenotype. Some suppression, but not enhancement, of the Age phenotype by daf-12(m20) was observed. In the case of the class 2 alleles, daf-12(m20) generally had no effect on the Age phenotype at 15° but enhanced it at 22.5°.

View larger version (79K): In this window In a new window Download PPT slide   Figure 6. Effect of daf-2 and daf-12(m20) mutations on median and maximum life span. Median and maximum life span is expressed as a percentage of the respective wild-type controls. Bars represent standard error. (A) Life span at 15°; (B) life span at 22.5°. Alleles are arranged in order of increasing severity of Daf-c, within each daf-2 class.

In a daf-2(+) genetic background, the daf-12 mutation reduced median and maximum life span at both temperatures (Table 3). At 15°, the weaker class 1 alleles e1365, m41, m577, and sa193 resulted in marginal increases in life span, whereas the strong class 1 allele m212 doubled both median and maximum life span (Figure 6A). At this temperature, the addition of daf-12(m20) resulted in a slight reduction in the median life span of the sa193 strain and a marked reduction in the median life spans of the e1365 and m212 strains. Significant shortening of life span was not seen in m41 and m577. At 22.5°, e1365, m41, m577, and sa193 resulted in large increases in life span, and addition of daf-12 resulted in some depression of the enhanced longevity (Figure 6B, Table 3). In daf-2(e1365); daf-12 strains, median and maximum life spans were depressed relative to e1365. In the daf-2(m577); daf-12 and daf-2(sa193); daf-12 strains, the extension of median life span was marginally reduced, and in the daf-2(m41); daf-12 strain, clear reduction of the extension of median (but not maximum) life span was seen. Conversely, the daf-12 mutation did not reduce the daf-2(m212) Age phenotype at 22.5°. Thus, daf-12(m20) appears to act as a weak suppressor of Age, at least with respect to median life span, more readily suppressing the weaker class 1 daf-2 alleles.

In the cases of the class 2 alleles e979, e1370, and e1391 at 15°, addition of daf-12(m20) had no significant effect on life span except for e979; daf-12, where maximum life span was enhanced (Figure 6A, Table 3). At 22.5°, addition of daf-12 generally enhanced median and maximum life spans (Figure 6B, Table 3). Addition of daf-12 increased e1391 median and maximum life span to 441% and 453% of N2, respectively, at 22.5°. One e1391; daf-12 animal lived to the age of 102 days.

Interactions between daf-2 and daf-12—Thermotolerance:
The effects of daf-12(m20) on thermotolerance did not parallel the effects on larval development and life extension. Although daf-12(m20) suppressed the class 1 Daf-c phenotype, and in some cases reduced class 1-enhanced longevity, daf-2(m41); daf-12 and daf-2(m577); daf-12 were more thermotolerant than the single mutants (percent increase in median survival 31.5 ± 0.5% and 27.8 ± 6.1%, respectively). Maximum survival was also enhanced (data not shown). Survival of daf-2(e1365); daf-12 was no different from that of e1365.

Class 2 alleles e1370 and e1391 interacted with daf-12(m20), resulting in arrested larval development at 25.5°; both double-mutant strains displayed increased maximum life spans at 22.5°. With regard to thermotolerance, the e1370 double mutants were slightly less tolerant than e1370, whereas the e1391 double mutants were slightly more tolerant (data not shown).

Interactions between daf-2 and daf-12—Internal hatching, late progeny, adult behavior, and morphology:
The late progeny and reduced motility traits were slightly enhanced by daf-12(m20). In combination with e979, e1370, e1391, and m212, daf-12(m20) resulted in a somewhat earlier onset of motility reduction, with full onset beginning 16, 4, 5, and 4 days after upshift of L4s to 22.5°, respectively. Conversely, death by internal hatching was strongly suppressed by daf-12(m20) in the class 2 alleles e1370 and e1391 but not e979 (data not shown). Internal hatching in the five class 1 alleles tested was unaffected by the presence of daf-12(m20). In all four cases, the presence of daf-12(m20) caused no clear enhancement or suppression of the adult body shrinkage or abnormal gonad phenotypes seen at 22.5° and 25.5°.

Of the eight daf-2 alleles tested in combination with daf-12(m20), three, e979, e1370, and e1391, resulted in late progeny at 22.5° and 25.5° in a daf-12(+) genetic background (Table 7). Addition of daf-12(m20) resulted in most cases in a marginal increase in the severity of the phenotype (data not shown). Thus, a very small number of late progeny were produced at 15° (0.15–0.4 late progeny per animal), whereas none were seen in the daf-2 single mutants at that temperature. An exception was e1391, where addition of daf-12(m20) reduced late progeny production at 22.5°, from 8.2 to 1.8 late progeny per worm.

The daf-2 null phenotype:
The deficiencies mDf11 and mDf12 fail to complement daf-2 (see MATERIALS AND METHODS). Although neither deletes the leftward or rightward markers, dpy-1 or unc-93, both deficiencies were shown by PCR analysis to lack daf-2 sequences encoding portions of the extracellular and protein kinase domains. Furthermore, mDf11/mDf12 animals exhibit an egg/L1 arrest phenotype, not the maternally rescued daf-2 null phenotype, indicating that the deficiencies include at least one essential gene in common, in addition to daf-2. Crosses of mDf12/qC1 males with mDf11/qC1 hermaphrodites, in which 88% of the adults were cross-progeny, 32% of the F₁ were egg/L1 lethal but there were no dauer progeny.

At 25.5° m65/mDf11 and m65/mDf12 animals formed dauer larvae nonconditionally without detectable embryonic or L1 arrest. However, the latter phenotype could have been maternally rescued. The relative severity of m65, mDf11, and mDf12 were compared by examining the Daf-c phenotype of each in trans to e1370 at 20°. At 20°, e1370 results in less than 1% dauer formation. In two trials, 4.7 ± 2.9% (N = 490) and 2.2 ± 0.7% (N = 1351) of m65/e1370 animals formed dauer larvae, whereas mDf11/e1370 and mDf12/e1370 resulted in 5.8 ± 5.8% (N = 579) and 13.0 ± 6.2% (N = 943) dauer formation, respectively. Thus, neither deficiency results in a phenotype significantly more severe than m65, supporting the view that m65 is a null [daf-2(0)] allele.

The daf-2(0) phenotype is complicated by the fact that th